environment collaborative environmental governance ...Örjan bodin* managing ecosystems is...

REVIEW SUMMARY◥

ENVIRONMENT

Collaborative environmentalgovernance: Achieving collectiveaction in social-ecological systemsÖrjan Bodin*

BACKGROUND: Current and future gener-ations are confronted with the complex taskof devising sustainable solutions to environ-mental problems. The coming decade mightdetermine whether humanity will be able toset a course toward a future of continued pros-perity on a planet whose ecosystems will deliverthe needed goods and services. A crucial pieceof this puzzle is achieving effective collaborationamong different public and private actors andstakeholders. Calls for solving environmentalproblems through collaborative governance em-phasize benefits from local to global scales—from artisanal fishermen avoiding the overfishingof local fish stocks by together agreeing uponsustainable practices, to states jointly committingto implement adequate measures to reducegreenhouse gas emissions. Although commonlyadvocated, achieving successful collaborationswhen confronted with complex environmentalproblems spanning geographical scales and

jurisdictional boundaries is an area wheresubstantial knowledge gaps remain.

ADVANCES: A growing amount of empirical ev-idence shows the effectiveness of actors engagedin different collaborative governance arrange-ments in addressing environmental problems.However, studies also show that actors some-times collaborate only as ameans of advocatingtheir own interests, while largely lacking a wil-lingness to contribute towards jointly negotiatedsolutions to common problems. Hence, col-laboration is sometimes unable to deliver anytangible outcomes, or merely produces sym-bolic outcomes such as aggregated wish listswhere conflicts of interest are left untouched.Clearly, no single blueprint exists for how to

succeed by using collaborative approaches tosolve environmental problems. One way ofapproaching this puzzle is through the lensesof the participating actors and the ways in

which they engage in collaboration with eachother. This approach entails directing atten-tion to who the actors are, what their interestsand motives are, who they collaborate with,and how the structures of such “collaborativenetworks” relate to the actors’ joint abilities toaddress different environmental problems.Emerging insights from recent research sug-

gest that the effectiveness of different collabora-tive network structures inaddressing environmentalproblemsdepends onhowthoseproblemsunfoldwithrespect to the followingcharacteristics: (i) varyinglevels of risk that actors

free-ride on others’ efforts; (ii) varying levels ofknowledge gaps, signifying different needs forsocial learning and deliberation among actorswith different backgrounds, experiences, andinterests; and (iii) whether these problems are,for all practical purposes, permanent or justtemporary.Also, long-standing research questions re-

garding whether governance structures thatare adequately aligned with ecosystem struc-tures and processes are more effective haverecently been addressed empirically. Early re-sults suggest several ways in which misalign-ments between the structure of a collaborativenetwork and the biophysical environment re-duce the ability to address environmental prob-lems effectively.

OUTLOOK: A more nuanced understandingof whether collaborative governance is themost effective way of solving environmentalproblems is needed. The capacity of collabora-tive governance to deliver sustainable solutionsfor any given environmental problem rangesfrom highly effective to essentially worthless.Future efforts must establish which factorsdetermine the exact location of any collabora-tive arrangement on this continuum.Emerging insights suggest that where a

collaborative arrangement falls on the spec-trum results from a complex interplay betweenseveral factors. The characteristics of theunderlying collective action problem are onefactor. Others are the characteristics of theunderlying biophysical system and how thesealign with the ways in which collaborative gov-ernance arrangements are constructed, institu-tionally embedded, andmanaged. Finally, thepatterns inwhich actors collaborate with eachother (or do not) is a factor that potentiallydetermines the effects that the other factorshave on a collaborative arrangement’s abilityto solve environmental problems.▪

RESEARCH

Bodin, Science 357, 659 (2017) 18 August 2017 1 of 1

Stockholm Resilience Centre, Stockholm University, 10691Stockholm, Sweden.*Corresponding author. Email: [email protected] this article as Ö. Bodin, Science 357, eaan1114 (2017).DOI: 10.1126/science.aan1114

Small-scale fishermen preparing their nets. Although collaborative approaches toenvironmental governance are increasingly advocated, a better understanding of if and howmultiactor collaboration in interlinked social-ecological systems is able to effectively addressvarious environmental problems is urgently needed.PH

OTO:NATUREPIC

TURELIBRARY/

ALA

MYSTOCKPHOTO/S

CIENCE

ON OUR WEBSITE◥

Read the full articleat http://dx.doi.org/10.1126/science.aan1114..................................................

on July 2, 2021

http://science.sciencemag.org/

Dow

nloaded from


REVIEW◥

ENVIRONMENT

Collaborative environmentalgovernance: Achieving collectiveaction in social-ecological systemsÖrjan Bodin*

Managing ecosystems is challenging because of the high number of stakeholders, thepermeability of man-made political and jurisdictional demarcations in relation to thetemporal and spatial extent of biophysical processes, and a limited understanding ofcomplex ecosystem and societal dynamics. Given these conditions, collaborativegovernance is commonly put forward as the preferred means of addressing environmentalproblems. Under this paradigm, a deeper understanding of if, when, and how collaborationis effective, and when other means of addressing environmental problems are bettersuited, is needed. Interdisciplinary research on collaborative networks demonstrates thatwhich actors get involved, with whom they collaborate, and in what ways they are tied tothe structures of the ecosystems have profound implications on actors’ abilities to addressdifferent types of environmental problems.

Ecosystems constitute complex entities span-ning geographical and temporal scales typi-cally not well-aligned with variousman-madejurisdictional and political demarcations.Hence, the ability to match the scale and

extent of ecosystems with appropriate structuresof governance suffers from institutional fragmen-tation (1). These considerations are at the heart ofthe research on institutional (or social-ecological)fit (2–6), and cross-border and cross-scale collab-oration is often seen here as a means by which toovercome such institutional fragmentation (7, 8).Furthermore, ecosystems are characterized by un-certainties and emergent behaviors (9). Therefore,developing better knowledge of ecosystem dy-namics through continual learning is consideredto be of key importance in environmental gov-ernance (1, 10). Collaboration is, in this realm, putforward as a means by which to (i) enhance thegeneration of newknowledge through social learn-ing (9, 11), (ii) better integrate important insightsfrom different knowledge systems (12), and (iii)diffuse knowledge and best practices among amultitude of actors (13). Also, governance of eco-systems involvesbalancingactors’different interests.If not, asymmetry of power and influence amongdifferent advocacy coalitions can, for example, leadtogovernance inertia, inhibitingeffectivemeasuresfor dealingwith environmental problems (14, 15).A common argument here is that collaborationacross opposing coalitions canhelp to unlock suchgoverning deadlocks (16, 17).Arguments in favor of multiactor collaboration

in addressing environmental problems are plen-tiful and stretch across many different fields of

research (8, 10, 18). This broad and multifacetedresearch uses different terminologies and per-spectives. The term “collaborative environmentalgovernance” is used here to capture collaborativeapproaches to environmental management in ageneral and inclusive sense. Although the argu-ments in favor of collaboration and the studiessupporting these claims are numerous, there isalso ample evidence that collaboration does notalways deliver substantial benefits. Hence, thereare reasons to caution against collaboration asan all-encompassing mode of government for allkinds of challenges (19). For example, it can bevery time consuming for a group of actors withdifferent backgrounds and interests to overcomeinitial collaborative barriers (20, 21). Hence, be-cause some pressing environmental problems callfor immediate action, mitigation through multi-actor collaboration might not always be the mostfeasible option. Further, in practice, actors oftendecide for themselves whom they wish to collab-orate with, what they want to accomplish, and inwhat types of collaborative venues (1, 22). Hence,governance through multiactor collaboration is,as compared tomore traditional and bureaucraticmodes of government, encumbered with criticalissues pertaining to various democratic qualitiessuch as transparency of decision-making proce-dures, legitimacy and accountability, and pro-cedural fairness (23). Managing collaborativeenvironmental governance initiatives thereforepresents public managers with novel leadershipchallenges (24, 25).The environmental issue of concern might be

so highly contested and riddled with issues ofasymmetries of power among the stakeholders,that hoping for collaboration as ameans of solvingenvironmental problems is quite simply naïve(26, 27). Studies of policy change have shown that

collaboration in highly contested policy issues doesnot necessarily have any substantive impact. Forexample, no substantial changes in Swiss nuclearenergy policies occurred during the years 2001through 2006, although the three opposing actorcoalitions did collaborate relatively intensively(15). Further, although the federally supportedgroundwater-management partnership in theVerde River Basin in the United States has in-stigated actors with conflicting interests to col-laborate in, e.g., sharing information, this has notlead to any substantial changes in values andbeliefs (28). This, in turn, has hindered the actorsfrom jointly taking any major steps toward gen-eratingmutually agreeable management options.Also, therehavebeen cases showing that striving

toward enhanced collaboration could in itself es-calate conflicts (29). Therefore, collaborative ini-tiatives that are unable to address conflicts ofinterest and deliberate in finding some form ofmiddle ground can fall short of producing any-thing other than a reinforcement of the currentstatus quo (30). Alternatively, they might fallshort of delivering anything other than a simplecompilation of the actors’ own wish lists or asimple agreement on vague and noncommittaldeclarations, largely concealing fundamentaltrade-offs and contradictions (31).The rapid uptake and rollout of collaborative

approaches to governance across different contexts(32, 33) has also created considerable uncertaintyand variability among actors with regard to whythey collaborate, what exactly they are supposedto (or want to) accomplish, and with whom [com-pare (34)]. This can result in actors spendingconsiderable time and resources on networking,leading to a high turnover of social ties, althoughthis does not necessarily lead to increased gov-ernance performance. For example, increasednetworking among planners engaged in en-hancing Swedish municipalities’ preparednessfor forthcoming natural disasters is seeminglynot associated with increased performance (35).The specific types of social ties actors develop

while engaging in collaboration also affect col-laborative outcomes. For example, social tiesutilized to merely exchange information can fa-cilitate social learning, albeit being ineffective inenabling any behavioral change (36), whereas socialties that build on deeper relations like friendshipcan facilitate such changes (37).

A integrated network-centric framework

Collaboration thus seems like an appealing andoften necessary, but not in itself sufficient, modusoperandi for addressing many of today’s environ-mental problems. Put bluntly, addressing theissue is clearly not as simple as just establishingcollaboration among a large set of actors andstakeholders, and then all will be well. Rather, thequestions are when and how collaboration iseffective, for what kind of environmental prob-lems is it useful, and if and how this relates tothe temporal and spatial characteristics of thegoverned ecosystems.One way of approaching this puzzle is through

the lenses of the participating actors and theways

RESEARCH

Bodin, Science 357, eaan1114 (2017) 18 August 2017 1 of 8

Stockholm Resilience Centre, Stockholm University, 10691Stockholm, Sweden.*Corresponding author. Email: [email protected]

on July 2, 2021


Dow

nloaded from


inwhich they engage in collaborationwith others.This entails directing attention to who the actorsare, with whom they collaborate, and how thestructures of such “collaborative networks” relateto the actors’ abilities to address different envi-ronmental problems (38, 39). Recent years haveseen a rapid increase in studies investigatingwhether various environmental problems haveinstigated the formationof collaborative networksand, if so, how these are formed. In addition, albeitto a lesser extent, studies have also investigatedhow these different forms affect the ability of theactors to address different types of environmentalproblems. This is by no means the only way inwhich to study collaborative environmental gov-ernance.However, it provides ameans of investigat-ing different collaborative processes through abottom-up approach (examining if, how, andwhyactors engage in different kinds of collaborationwith certain others), while simultaneously provid-ing an analytical vehicle for investigating collab-orative performance at the group level (examiningrelationshipsbetweendifferent collaborativenetworkcharacteristics and collective abilities to solveenvironmental problems). Hence, a collaborativenetwork perspective thus constitutes a frameworkthat enables the cross-fertilizationof insights acrossdifferent studies and fields of research. Networkscan be characterized in numerous ways, althoughcharacteristics often focused on are (i) degree ofnetwork cohesiveness (e.g., density of relations);(ii) degree of network centralization (the extentto which one or a few actors act as hubs); (iii)degree of network fragmentation (i.e., if and towhat extent the network consists of different sub-groups); and (iv) degree of connectivity acrossdifferent types of actors (i.e., homophily andheterophily) (Fig. 1).

A key factor that distinguishes environmentalproblems from many other collective action prob-lems in general is that environmental problemsare inevitably tied to the complex structures andprocesses of boundary-spanning ecosystems. Thus,effective and long-lasting solutions to environ-mental problems require these ecosystem charac-teristics to be explicitly taken into account (9, 40).However, it is not uncommon that studies of col-laborative environmental governance are entirelyfocused on the social and political processes, andthe specifics of the ecosystem as the target for thecollaborative efforts are largely disregarded. Thisis not to say that these studies are missing thepoint. On the contrary, these processes are ofcrucial importance for any kind of collaborativeundertaking. However, these studies do not inves-tigate if and in what ways the specific biophysicalcharacteristics of the ecosystems pose any con-straints with regard to how collaborative arrange-ments should ideally be devised. Hence, takingstock in the large body of research on social-ecological fit that argues that there is no one-size-fits-all governance arrangement that works wellacross all possible social-ecological contexts (2–6, 41),it seems crucially important to advance understand-ing regarding howwell a collaborative arrangement“fits” to the specifics of the environmental problembeing addressed. Recent theoretical and methodo-logical innovations in multilevel network analyseshave made headway in facilitating interdisciplinaryinquiries where both social and ecological dimen-sions of collaborative environmental governanceare analyzed together (42–45). Therefore, an explicitnetwork perspective on collaborative environmentalgovernance can be used as an integrated frameworkin investigating which social structures and pro-cesses are conducive to addressing which kinds

of environmental problems, and in which kindsof social-ecological contexts.

Fit to the collective action problemCollaborative learning

Many, if not most, environmental problems canbe characterized as collective action problems.However, collective action problems come indifferent shades. A key argument in favor ofcollaboration is how it facilitates learning (10).Learning is here conceived as a collective actionproblem where processes that involve sharingexperiences and engaging in collective deliberationare in focus [social learning; see, e.g., (11)]. For suchprocesses to materialize, actors need be sociallylinked with others in suitable ways. Learningabout complex problems typically requires theactors to draw from a range of knowledge do-mains and expertise, which differs substantiallyfrom learning about problems that are well con-fined within a specific knowledge domain (com-pare inter- versus intradisciplinary research).Addressing complex problems is benefited bythe coming together of actors with different edu-cational backgrounds, roles, and occupations; there-fore, a strong tendency of similar actors flockingtogether in isolated subgroups could be detri-mental (compare Fig. 1C). Actors who only inter-act within their own subgroups easily developtheir own subcultures with a sense of “us andthem,” and different and often incompatibleperceptions of the problems at hand and how tobest solve them emerge between the subgroups(14, 15, 46–48). For example, it has been shownthat limited interaction between subgroups of tunafishers has suppressed collective learning, whichhas led to suboptimal harvesting practices (49).By contrast, a study of collaborative coastal-zone


Fig. 1. Different structural characteristics of networks. (A) Representa-tion of a cohesive collaborative network comprising numerous collaborativeties between actors engaged in coastal-zone management in Sweden (50).Thedifferences in centrality between the actorswith themost connections andthose with an average number of connections are relatively small, and theclosed triangular structure (inset) is a common building block in this network(two friends of a common friend also tend to be friends).The centralizationscore is 0.26 (on a scale from 0 to 1), and the modularity index that capturesthe extent to which the network consists of subgroups peaks at 0.07 (on ascale from 0 to 1). (B) Representation of a centralized collaborative networkfrom a United Nations Educational, Scientific, and Cultural Organization(UNESCO) biosphere reserve in Canada (100), where the differences in

centrality between the actors with the most connections and those with anaverage number of connections differ substantially.The open triangularstructure (inset) is a commonbuilding block in this network (an actor connectstwo unconnected actors).The centralization score is 0.63 and the modularityindex is ~0. (C) Representation of a more compartmentalized collaborativenetwork of small-scale fishermen in east Africa (70).The colors representfishermen using different fishing gear (traps, nets, etc.), and the dotted linesenclose different identified cohesive subgroups (subgroup membership alsodesignated by symbol shape). The subgroups partly coincide with geartype. The building block capturing two socially connected actors using thesame gear (inset) is common in this network. The centralization score is0.11, and the modularity index peaks at 0.58.

RESEARCH | REVIEWon July 2, 2021


Dow

nloaded from


management suggests that collaborative networks,in which heterogeneous actors are not confinedto isolated subgroups only consisting of theirimmediate peers, facilitate learning about complexproblems such as how to accomplish ecosystem-based management (50). However, it was alsoindicated that a similar learning effect couldbe accomplished through facilitation by actorsoccupying central positions in the collabora-tive networks.Learning about problems that are less com-

plex, typically confined within a specific knowl-edge domain, does not necessarily benefit frombringing together a heterogeneous set of actors.Instead, here it is often more relevant to framethe learning process as a process of diffusion. Thestudy of specific structural characteristics of socialnetworks well suited for diffusion constitutes aresearch field by itself (51), although the positiveeffect of collaborative networks characterizedby high densities of social ties for the spread ofnew management practices in environmentalgovernance has been empirically demonstra-ted (37, 52).A key component in addressingmany environ-

mental issues is the ability to innovate new solu-tions to sometimes old problems (53). Innovationcan be framed as a result of learning, althoughwith more emphasis on learning that favors de-liberation and thinking outside the box. Thislargely resembles the challenge of addressingcomplex problems. However, because novelty,in part, implies breakingwith current establishednorms and perceptions, an overly cohesive col-laborative network could contribute to the re-inforcement of current perceptions thus makingitmore difficult for new ideas to emerge and findsupport (54). This has been demonstrated amongfarmers in Australia, where those who instigatedmore transformative farming practices werepart of further-reaching but sparser collaborativenetworks as compared to those who were moreprone to incremental changes (55). This alsotouches upon the classic work by Granovetter inwhich he showed that far-reaching and weaksocial ties are more important when people areseeking novel information (56).

Coordination or cooperation

Although learning is of crucial importance ingoverning complex ecosystems, it is what theactors actually do that matters to the environ-ment. Many collective action problems can bedivided into two broad classes—coordinationversus cooperation problems. The former describesa situation where all or most actors agree on whatthey want to accomplish, and getting there is morea matter of orchestrating the actors’ differentactivities in efficient ways (57). Joint efforts toeradicate an invasive salt-marsh cordgrass spe-cies in the San Francisco Bay in California servesas an example of a coordination problem (58).The latter corresponds to problems where actorsdisplay different opinions and interests and whereproblem solving would, by necessity, involve nego-tiations and deliberations to reach common agree-ments. Often this implies that actors will have to

retract a bit from what they would ideally preferin terms of, for example, resource utilization. Aspecial class of cooperation problem is when thereare inherent trade-offs, which can be framed as a“distribution” problem (1). Accomplishing sustain-able harvest levels in multinational high-sea fish-eries, where multiple actors compete for a limitedresource, serves as an example of a cooperativedistribution problem (17). Further, coordinationand cooperation problems have been framed aslow-risk and high-risk, respectively (59). Risk wasoriginally framed as the risk of actors defectingbut has evolved to a broader conceptualizationof risks in collaborative endeavors (60). Several

empirical studies support the notion that co-herent and dense collaborative network struc-tures are better at addressing high-risk cooperationproblems, whereas more centralized and sparsenetworks are better for low-risk coordinationproblems. More specifically, a network conduciveto managing cooperation problems is charac-terized by actors tending to reciprocate incomingsocial ties and form triadic structures (two friendsof a friend will also be friends; see Fig. 1A). Thesedense structures help exert social pressure tocomply, but they also help develop mutual trust.A centralized network is characterized by moreopen structures, i.e., two friends of a friend willnot necessarily also be friends, and by someactors being much more connected (central) thanothers (57, 59) (Fig. 1B). These open and sparserstructures facilitate coordination without neces-sitating that actors invest lots of resources inupholding a relatively high number of social ties.These network characteristics are also conceived

as representing bonding versus bridging socialcapital, respectively (22).

Temporal or long-term problems

Often, collaborative governance arrangements areinitiated to address long-termenvironmental prob-lems, such as climate change (61). Unless suchcollaborative processes are providedwith fundingand support over substantial time frames, theywill struggle to accomplish desired results (20).If, however, collaborative networks are sustainedover time, they can lead to the cultivation andmain-tenance of commonnormsand routinedeliberation(62), which are key factors in addressing long-termenvironmental problems (18).Transient environmental problems, however,

such as eradicating a specific invasive species orstopping an escalating wildfire, require a rapidresponse. Thus, they might be better addressedthrough a rapid mobilization of relevant actorsin ad hoc collaborative networks. Furthermore,because time is often scarce, accomplishing effec-tive coordination is of utmost importance. There-fore, more-centralized networks, where somespecific actors act as the spiders in the web bydistributing and coordinating tasks, are favorable(58) (Fig. 1B). On the other hand, such networkstructures are less suited to addressing cooperationproblems (59). Accordingly, unless the collectiveaction problem itself is only about coordinatingactors already agreeing onwhat needs to be done,centralized ad hoc networks will be more effec-tive if they are drawn from underlying andmore-permanent collaborative networks, wheremutualtrust and willingness to comply is already wellestablished (62, 63). This illustrates an interplaybetween the formation of effective and centralizedadhocnetworks andunderlying, dense, and longer-lasting collaborative networks.

Fit to the ecological context

Not only should a collaborative network fit thespecifics of the collective action problem, it shouldalso fit underlying biophysical characteristics. Ona conceptual level, social-ecological fit impliesthat the structure of a collaborative network(the actors and their collaborative ties) shouldbe aligned with the structures of the biophysical(ecological) system being governed. However, toadvance such a blanket statement, there is a needto more precisely define what would be a favor-able fit, and why. This involves addressing twobroad questions—namely, who should ideally beinvolved in a collaborative network and withwhom should they ideally collaborate? Appropriateanswers to these questions are, from the perspec-tive of social-ecological fit, inherently related to thecharacteristics of the underlying biophysical sys-tem. Several recent and complementary network-centric frameworks facilitate answering thesequestions (42–45). These frameworks departfrom a multilevel-network approach, where thesocial and the ecological systems are representedas separate but interconnected network layers. Thesocial-network layer consists of actors and theirrelationships, and the ecological-network layerdescribes the ecosystem as sets of interdependent


Box 1. Management challenges in colla-borative environmental governance.

• How to create and maintain collaborativenetworks that are able to address toughproblems involving deep-rooted conflicts ofinterest, while simultaneously being condu-cive to the efficient coordination of relativelysimple tasks

• How to facilitate social-tie formation pro-cesses in the local context such that the evolv-ing collaborative network develops desirableglobal structural properties, including a goodfit to the biophysical context

• How to best engage actors in collaborativenetworks even though some of them are notinterested, are interested for the “wrong”reasons, or use the collaborative venue onlyas a way of obstructing any changes to thestatus quo

• How to create and maintain collaborativenetworks that are flexible and adaptable tochanges, yet stable enough to facilitate thedevelopment of mutual trust and sharedcommitment



Dow

nloaded from


ecological components (Fig. 2). By using such asocial-ecological network representation of a col-laborative environmental governance arrangement,it is possible to distinguish two dimensions ofsocial-ecological fit, namely, “horizontal” and“vertical” fit. The former is concerned with howwell social and ecological network ties are alignedacross the layers, whereas the latter is concernedwith how the different social and ecological layersare interconnected.

Horizontal fit

Ecosystems consist of interdependent compo-nents. These ecological interdependencies arefundamental to the functioning of ecosystems,and compromising ecological connectivity willthreaten the ability of ecosystems to provide theecosystem services that societies are relying on(9, 64). Hence, the maintenance of these links iscrucial. However, this is often a challenging taskwhen human use of natural resources increases.This becomes particularly challenging if any twointerdependent ecological components are man-aged by different actors who are not coordinat-ing their managing activities. An example of suchmismanagement would be when two actors whoare each managing separate forest patches fail tomanage their lands in a way that facilitates spe-cies dispersals, which could threaten a commonforest-dwelling metapopulation whose viabilitydepends on its ability to freely relocate betweenthe patches (65). Studies of Balinese farmers, onthe other hand, demonstrate how they, throughcollaboration, collectively reduce the spread ofpests across their ecologically interconnected ricefields by synchronizing their water use (66). Thus,this implies that a better social-ecological fit isaccomplished if links in the ecological networkare paired with links in the collaborative network.This can be described by using the notion of social-ecological building blocks (see Fig. 3A). A social-ecological building block represents a minimalset of actors and ecological components, and theirdifferent types of interdependencies (links), thatdescribes a theoretically important configurationof actors and ecological resources. Analogous toregression analysis, by using multilevel exponen-tial random graph models, it is furthermore pos-sible to statistically infer if and to what extentdifferent building blocks explain empirically ob-served structural characteristics of social-ecological(multilevel) networks (43, 67).A similar argument based on alignment can

be applied to cases when two actors are managing(or competing for) the same ecological component(Fig. 3B). In such cases, the utility of collaborationis even more pronounced, especially if both actorsare using the shared component for extractivepurposes. In this type of setting, it might be ratio-nal for the actors to extract as much of the re-source as they can to safeguard themselves frombeing left with nothing if the other actor was tomaximize its extractions. This typically leads tooverharvesting and resource depletion, unlessresource extraction is strictly regulated and en-forced by a third party (e.g., public authorities)and/or the resources are privatized (68). Such

measures are often neither practically feasiblenor even desired. Hence, if the actors are to avoiddepleting their common resource, they need tocollaborate in order to devise and enforce com-monly agreed upon regulations and harvestingpractices (18, 69).These theoretically derived arguments in favor

of certain social-ecological building blocks haverecently been exposed to empirical case studies.Even though this interdisciplinary research is stillin its infancy, some insights are starting to emerge.Results from studies of a large-scale biodiversityconservation initiative in Australia and a small-scale fishery in east Africa indicate that collab-orative networks where actors with stakes incommon ecological components tend to collab-orate (Fig. 3B) are associated with better preser-vation of ecological resources and more effectivemanagement (70, 71). Results are, so far, lessconclusive when it comes to the building blockencapsulating the alignment of social and ecolog-ical connectivity (Fig. 3A). Case studies rangingfrom local scales, such as intermunicipality col-laboration on wetland management, to globalscales, such as species dispersals across the terri-tories of states, suggest that actors do not collab-oratewith others in themanagement of ecologicallyinterconnected resources more than would beexpected by chance (43, 71–73). Reasons for this

could range fromlegal obstacles that prevent actorsfrom collaborating across jurisdictions (7) to a lackof comprehension in regard to the existence ofecological interdependencies (74). Furthermore,more empirical inquiries into whether the align-ment of social and ecological connectivity wouldlead to more desirable ecological outcomes areneeded.

Vertical fit

A well-fitting collaborative network does not onlyentail aligning patterns of social and ecologicalconnectivity. As stated, ecosystems should ideallybe managed as systems and not as sets of isolatedcomponents; hence, the patterns in which theactors are tied to the ecological components areof crucial importance. A social-ecological build-ing block representation of vertical misfit is whenan actor only manages a fraction of the eco-system, i.e., just one of two interconnected eco-logical components (Fig. 3C). An example of suchmisfit is when landscapes are divided into differ-ent administrative, ownership, or managementcategories. This implies that different actors willbe in charge of different categories, although thecategories themselves are merely capturing differ-ent components and/or aspects of the coherentlandscape (therefore, they are ecologically inter-dependent). Nonetheless, this division of the


Fig. 2. A social-ecological network model of an integrated social-ecological system. Themultilevel network–modeling approach is illustrated with a stylized small-scale fishery system, whereactors are represented by fishing vessels (social nodes), and ecological components are representedby different targeted marine species (ecological nodes). The red links represent collaborative ties,the blue links represent trophic interactions among the marine species, and the black links showwhich vessel is targeting which marine species (these vertical links thus capture how differentactors have different stakes in different components of the ecosystem). This approach can be usedto model other systems. For example, the social nodes could constitute individuals, groups,organizations, or any other abstraction of an actor or governing entity, and the ecological nodescould constitute other biophysical entities such as habitat patches or more abstract ecologicalentities (45) such as ecosystem services (44).



Dow

nloaded from


landscape into different components is likely toexacerbate habitat change and thus the fragmen-tation of contiguous land covers (75).A tighter feedback loop between an actor’s

managing activities and whatever environmentaloutcomes these activities give rise to on an ecosystemlevel is, however, accomplished if the actor islinked to both components (Fig. 3C). This is alsoreferred to as scale matching (5). In economicterms, this implies that potential ecological exter-nalities have been internalized (70). Emerginginsights suggest that this building block is morecommon in collaborative networks that are per-forming reasonably well (70, 71), although empir-ical research investigating whether collaborativenetworks experience this type of vertical fit is stillvery scarce.Actors are often situated at different adminis-

trative levels. These levels typically correspond todifferent geographical scales (compare local re-source extractors and regional managers). Manyecological processes interact across scales; there-fore, social ties linking actors across these ad-ministrative levels imply a better alignment ofcollaborative structures and ecological cross-scaleinterdependencies (42). Hence, vertical cross-levelsocial ties indirectly linking actors at the sameadministrative levels could enhance horizontalsocial-ecological alignment (Fig. 3D). In a studyof collaborative intermunicipal wetlandmanage-ment, it was found that a variety of social-ecologicalbuilding blocks resembling the idea of coordinatingactors indirectly linking two other actorswere over-represented in the collaborative network, suggest-ing that actors have a propensity to engage incollaborative structures where coordination isfacilitated through a third party (76). Further, astudy of estuary-watershed governance indicatedthat actors’ perceptions of the productivity ofsocial ties linking local and regional levels werenegatively affected if these ties were confined toa very limited set of central actors, although thestudy also revealed that such an effect was inter-twined with other network effects (42). However,the arguments behind the presumed benefits oflinking levels are not limited to studies that usean explicit social-ecological network representa-tion of actors and ecological components. Thecore arguments presented here largely resemblesome of the presumed benefits of polycentricgovernance (77), and the utility of scale-crossingand multilevel environmental governance moregenerally (6, 78, 79). Thus, the theoretical argu-ments and the empirical evidence supporting thepresumed benefits of cross-level collaborative tiesis quite substantial.

Compounded environmental problems

Environmental problems are often best describedas aggregates of more or less interdependent sub-problems [compare (80)] hence simultaneouslydisplaying a range of collective action–problemcharacteristics. For example, even though cooper-ation and coordination seemingly benefit fromrather different network structures (Fig. 1, Aand B), empirical collaborative networks tend todisplay both types of structures (50, 57, 60, 81, 82).

Thus, it appears that collaborative networks areoften formed in response to both types of collectiveaction problems, although it should also be fac-tored in that actors do not exclusively create socialties on the basis of the nature of the collectiveaction problem [compare (22)]. Furthermore, theproblem specifics of an environmental issue willlikely change over time. This implies that whatmight constitute an effective collaborative networkshould also change over time (39). Among dairyfarmers in the eastern United States, the buildupof weak social ties was integral in the enablementof a transformation to new and novel farmingpractices; however, the farmers did not maintainthese weak relationships after the practices hadtransformed (83). Hence, after the transformation,these weak ties were likely no longer needed. Alongitudinal study of climate change–mitigationpolicy development in Switzerland further dem-onstrated that the policy networks changednotably between the decision-making and theimplementation phases (84).Taken together, all of this suggests that multi-

purpose collaborative networks that are able toaddress a range of collective action problems, andthat can adapt to changes in the nature of theseproblems, are better suited to addressing environ-mental problems. Less known, however, is if beingfit to various collective action problems and beingfit to the ecological context constitute two inde-pendent dimensions of fit. Claims conceptually

favoring interdependency abound, signified by theestablishment of several widely used frameworksemphasizing the need for integrated social-ecologicalsystems perspectives [e.g., (10)]. Ongoing attemptsto mitigate climate change serve as an illustration.Climate change mitigation appears to struggle incomparison with, for example, the success of themultilateral treaty that swiftly reduced the emis-sion of substances that deplete the ozone layer.Both these environmental problems are similar inthat they engage many states in tough negotia-tions. Nonetheless, they deviate in performance(61). Such deviations are regularly attributed tocontextual social-ecological differences. A social-ecological network perspective can help disentanglesome of these social-ecological contextual differ-ences into clearly articulated, theoretically grounded,and measurable characteristics, specifying waysin which actors and the environment are entangled.Bringing the social-ecological and collaborativenetwork perspectives together in a unifying frame-work could therefore facilitate integrated studieswhere “classic” collective action problems (learning,coordination, cooperation, etc.) and social-ecologicalfit are analyzed together.An important question for further research

should be if and how the utility of collaborativenetwork structures conducive for solving coordi-nation problems (Fig. 1B) depends on how socialand ecological connectivity is aligned horizontallyand vertically (Fig. 3). Such research endeavorswouldcontribute inunpacking the social-ecologicalcontext and more precisely investigate potentialcausal pathways in which social-ecological inter-dependencies influence actors’ abilities to solvecollective action problems of various kinds. Astudy of the relatively effective management re-sponse that followed the establishment of theinvasive Indo-Pacific lionfish (Pterois miles andPterois volitans) across a set of marine protectedareas (MPAs) in Jamaica serves as an illustrationof research pointing in that direction (85). Be-cause of the high ecological connectivity amongtheMPAs, an effective response required all MPAmanagers to apply adequate eradicative mea-sures to their sites simultaneously. Hence, therewas no need to adhere to any specific sequence oferadiation across the MPA sites. Devising andimplementing a sequential response among alarge set of local MPA managers would likelyhave required more coordination effort than justagreeing on a common starting time. Hence, theneed for thorough horizontal coordination amongthe managers was lowered. The study further-more suggests that the managers’ synchronizedresponses were made possible largely becauseof a high level of cross-level connectivity, i.e.,the local managers were well connected withhigher-level authorities that coordinated theirresponse. This corresponds to the social-ecologicalbuilding block where a mediating actor facilitatescollaboration between any two actors managingtwo interconnected ecological components (Fig.3D). Therefore, because of the high ecological con-nectivity, vertical cross-level coordination seem-ingly became more important than horizontalcoordination (Fig. 1B, inset) in enabling an


Fig. 3. Social-ecological building blocks.(A and B) Horizontal fit, i.e., alignment of socialand ecological connectivity. (A) To the left, twoactors (red) managing two separate but inter-connected ecological components (green) arenot collaborating, whereas to the right they are.(B) To the left, two actors managing the sameecological component do not collaborate,whereas to the right they do. (C and D) Verticalfit across different network layers. (C) To theleft, the actor is managing one of two inter-connected ecological components, whereas tothe right the actor is managing both compo-nents, thus internalizing ecological externalities(closing the social-ecological loop). (D) Tothe left, the actors managing interconnectedcomponents are not collaborating [as in (A), left]and only one of them is collaborating with thepotentially mediating actor operating on a higheradministrative level (orange). To the right, thevertical cross-level social ties of the mediatingactor indirectly connect the two other actors.



Dow

nloaded from


effective response to this specific environmen-tal problem.

Collaborative networks and leadership

Collaborative networks are made up of actorswith different capabilities, interests, and intentions.Hence, the effectiveness of collaborative environ-mental governance in addressing environmentalproblems can only partly be understood from astructural collaborative network perspective. Forexample, a highly centralized network conduciveto efficient coordination (Fig. 1B) might still failif the centrally located coordinator is not doinghis or her job. Thus, the effectiveness of a col-laborative network results from the interplay be-tween the overall structure of the network, thecharacteristics of its actors, and the networkpositions that they occupy. Studies of small-scale fisheries have shown that the utility ofcoherent collaborative networks conducive forcooperation is amplified if suitable leadership isin place (86, 87). Recent research further suggeststhat effective collaborative environmental gov-ernance requires a range of different leadershipqualities [compare (24)] Below, some leadershipqualities that are strongly related to the struc-ture and functioning of collaborative networksare discussed.

Network positions andleadership qualities

The crucial importance of spanning boundaries(also referred to as bridging or brokerage) isemphasized in research and practice. A bound-ary spanner connects different types of actors,and/or organizational and biophysical levels andscales (compare to vertical social-ecological fitdiscussed earlier) that would otherwise be dis-connected or only weakly connected (88). Innetwork terms, a boundary spanner occupiesa position in between many others, spanningstructural holes in the network (88). Leader-ship executed by boundary spanners has beenshown to increase mutual trust (16, 17), and tohelp build adequate support in attempts toaddress environmental problems through far-reaching transformational changes in manage-ment and perceptions (24, 89). However, it hasalso been demonstrated that boundary spannersmight utilize their position mostly for personalbenefit (88); they might hold certain perceptionsand attitudes that can impede success in col-laborative endeavors (28), or, although they maycontribute positively to collaborative outcomes,they themselves might be penalized (90).Central actors, i.e., the ones that have con-

siderably more social ties than others, are wellsituated to execute leadership that facilitates col-lective action. Their central position not only in-cludes facilitating coordination of activities butalso synthesizing others’ insights and perceptionsto enable collective sense making [e.g., (24)] andthe diffusion of new ideas and practices (51).Moreover, central actors occupy a position wellsuited for helping bridge across different bounda-ries (through the sheer number of ties, which isnot the same as occupying a boundary-spanning

position) (50, 91). However, as previously stated,these presumed benefits that derive from occupy-ing a central position are inherently tied to thecentrally positioned actors’ leadership skills, howthey are perceived by others, and which resourcesthey have at their disposal. Cognitive limitations,for example, pose constraints on the amount ofcoordination that can effectively be carried out(92). It has been demonstrated that appointing(and funding) a specifically designated coordinator,a network administrative organization, can beinstrumental in realizing the potential benefitsthat occupying a central position in the collab-orative network can bring about (17, 21, 58, 93).Furthermore, it has been suggested that effec-tive coordination is benefited by the ability of thecentral actors to exert some pressure on others tocomply; hence, it is beneficial if they possess someauthoritative capacities (57).Studies have shown that some specific actors

who are engaged in collaborative endeavors actas “risk mediators” in that they tend to occupynetwork positions associated with tight bondingstructures (60) (Fig. 1A, inset). Thus, managingrisks in collaborative undertakings can be thoughtof as a division of laborwhere some actors executeleadership specifically intended to mediate riskyrelationships, thereby enabling others to allocatemore attention to less risk-prone collaborative en-deavors (e.g., coordination). For example, a studyof collaborative urban-development planning re-vealed that state agencies did most of the “heavylifting” in managing risks in various collaborativerelationships (60).

Network weaving

Collaborative networks, like other social networks,are not static; they continually evolve as actorsadjust to different endogenous and exogenousdrivers of change.Hence, different network struc-tures do not emerge by chance, nor are positionswithin the network distributed randomly. Develop-ing a better understanding of collaborative networkdynamics thus involves identifying themechanismsthat make certain actors engage in collaborationwith certain others, as well as identifying what itis that makes it more or less attractive to engagewith certain actors [e.g., (84)]. This touches uponyet anotherdimensionof leadership in collaborativeendeavors, i.e., how leaders directly or indirectly en-gage in creating, intervening in, and shaping net-works [“network weaving,” see (43, 89, 94, 95)].The formation of network ties can, for exam-ple, be stimulated through direct engagement,or through establishing collaborative venues(often referred to as collaborative institutions).The former is about engaging directly with otheractors, potentially through brokering. The latteris about convening the formation of ties throughthe establishment of collaborative venues whereactors are invited to work together to addresscertain predefined issues and problems (1). Actorsmight bemandated to participate in these venues,or participation could be voluntary. Dependingon the context, the size of a venuecould range fromvillage meetings where local fishermen gatherto discuss fishing practices to multinational col-

laborative platforms such as the IntergovernmentalPanel on Climate Change. Contemporary environ-mental governance systemsareoften characterizedby a high number of venues [e.g., (96)]. For ex-ample, a study ofwater governance inSanFranciscoBay, California, showed that the number of actorsand venues were bothmeasured in the hundreds(81). In such settings, largely resembling polycen-tric governance systems where decision-making isdistributed acrossmultiple fora, the collaborativearenas confronting actors and stakeholders arenot only made up of many other actors and theirsocial ties, but also span multiple venues andmultiple policy issues; all are potentially inter-dependent in complex ways (1, 97). How thiscomplex “ecology of games” affects collaborativeenvironmental governancehas stirredup scholarlyinterest. Early findings suggest that the morevenues that individual actors participate in, thehigher they tend to perceive venue effectivenessand the amount of resources they can derive fromvenueparticipation (98,99), althoughbroad venueparticipation can also negatively influence policysatisfaction (1).

Conclusion and outlook

Much is known about collaborative networksand how they tend to be formed and shaped.However, merely establishing a collaborativenetwork in no way guarantees that environ-mental problems will be effectively addressed.Future efforts are needed to determine whenand in what contexts collaborative approachesare most effective, and when other approachesto solve environmental problems are better suited.The path forward involves addressing a range

of critical research questions. Our understandingof how certain collaborative network structurescontribute to different governance outcomes, andhow they interact with different aspects of agencyand leadership, is, at present, often vague and/or lacking firm empirical evidence. Further, re-search on how collaborative governance arrange-ments are more or less well fitted to variouscharacteristics of the ecosystems, and what thisimplies for governance outcomes, is very scarce.In particular, if and how being fit to the specificsof the collective action problem and being fit tothe ecological context interact is largely unchartedterritory. Assessing such entangled causal relation-ships between collaborative environmental govern-ance, social-ecological fit, andgovernanceoutcomesrequires further advancements of contemporaryinterdisciplinary theories and methods.Environmental problems are often composed

of a series of different kinds of interdependentcollective action problems. However, more effortsare needed if we are to understand if and howcollaborative networks that encompass a match-ing set of desirable structural characteristics con-ducive to addressing this range of problems canbe created and maintained. To be both sociallyand ecologically fit to the environmental problemsat hand, such multifunctional and multipurposecollaborative networks would need to strike afavorable balance betweenmany ideal, and oftencontradicting, structural characteristics. This calls




Dow

nloaded from


for further efforts to advanceunconventional formsof public and private leadershipmore focused onnetwork weaving and facilitation, and less oncommand and control (Box 1). Furthermore,collaborative governance initiatives are oftenestablished as projects, with funding for a limitedtime (20). The underlying environmental prob-lems, however, are often more enduring; hence,a fundamental challenge is to better understandhow collaborative endeavors can be better adoptedby formal bureaucracies and incorporated intoexisting government structures and processes.Many of the most pressing and complex envi-

ronmental problems of today operate at regionaland global scales. Furthermore, instigating andmaintaining effective collaboration might be theonly feasible option to address environmentalproblems at these scales. A substantial part ofcurrent research of collaborative networks in envi-ronmental governance is, however, conducted onsmaller scales. This suggests that more researchefforts should be directed toward the regionaland global scales.

REFERENCES AND NOTES

1. M. Lubell, Governing institutional complexity: The ecology ofgames framework. Policy Stud. J. 41, 537–559 (2013).doi: 10.1111/psj.12028

2. O. R. Young, The Institutional Dimensions of EnvironmentalChange: Fit, Interplay, and Scale (MIT Press, 2002).

3. V. Galaz, T. Hahn, P. Olsson, C. Folke, U. Svedin, inInstitutions and Environmental Change: Principal Findings,Applications, and Research Frontiers, O. R. Young, H.Schroeder, L. A. King, Eds. (MIT Press, 2008), pp. 147–186.

4. G. Epstein et al., Institutional fit and the sustainability ofsocial–ecological systems. Curr. Opin. Environ. Sustain. 14,34–40 (2015). doi: 10.1016/j.cosust.2015.03.005

5. G. S. Cumming, D. H. M. Cumming, C. L. Redman, Scalemismatches in social-ecological systems: Causes,consequences, and solutions. Ecol. Soc. 11, 14 (2006).doi: 10.5751/ES-01569-110114

6. J. Edelenbos, I. van Meerkerk, Connective capacity in watergovernance practices: The meaning of trust and boundaryspanning for integrated performance. Curr. Opin. Environ.Sustain. 12, 25–29 (2015). doi: 10.1016/j.cosust.2014.08.009

7. B. A. Cosens, Legitimacy, adaptation, and resilience inecosystem management. Ecol. Soc. 18, 3 (2013).doi: 10.5751/ES-05093-180103

8. B. Walker et al., Looming global-scale failures and missinginstitutions. Science 325, 1345–1346 (2009). doi: 10.1126/science.1175325; pmid: 19745137

9. N. L. Christensen et al., The report of the EcologicalSociety of America Committee on the Scientific Basis forEcosystem Management. Ecol. Appl. 6, 665–691 (1996).doi: 10.2307/2269460

10. C. Folke, T. Hahn, P. Olsson, J. Norberg, Adaptive governanceof social-ecological systems. Annu. Rev. Environ. Resour. 30,441–473 (2005). doi: 10.1146/annurev.energy.30.050504.144511

11. M. S. Reed et al., What is social learning? Ecol. Soc. 15, r1(2010). doi: 10.5751/ES-03564-1504r01

12. M. Tengö, E. S. Brondizio, T. Elmqvist, P. Malmer,M. Spierenburg, Connecting diverse knowledge systems forenhanced ecosystem governance: The multiple evidence baseapproach. Ambio 43, 579–591 (2014). doi: 10.1007/s13280-014-0501-3; pmid: 24659474

13. P. Matouš, Y. Todo, D. Mojo, Roles of extension and ethno-religious networks in acceptance of resource-conservingagriculture among Ethiopian farmers. Int. J. Agric. Sustain. 11,301–316 (2012). doi: 10.1080/14735903.2012.751701

14. P. A. Sabatier, An advocacy coalition framework of policychange and the role of policy-oriented learning therein.Policy Sci. 21, 129–168 (1988). doi: 10.1007/BF00136406

15. M. Fischer, Coalition structures and policy change in aconsensus democracy. Policy Stud. J. 42, 344–366 (2014).doi: 10.1111/psj.12064

16. I. van Meerkerk, J. Edelenbos, The effects of boundaryspanners on trust and performance of urban governancenetworks: Findings from survey research on urbandevelopment projects in the Netherlands. Policy Sci. 47,3–24 (2014). doi: 10.1007/s11077-013-9181-2

17. H. Österblom, Ö. Bodin, Global cooperation among diverseorganizations to reduce illegal fishing in the SouthernOcean. Conserv. Biol. 26, 638–648 (2012). doi: 10.1111/j.1523-1739.2012.01850.x; pmid: 22624623

18. T. Dietz, E. Ostrom, P. C. Stern, The struggle to govern thecommons. Science 302, 1907–1912 (2003). doi: 10.1126/science.1091015; pmid: 14671286

19. T. M. Koontz, C. W. Thomas, What do we know and need toknow about the environmental outcomes of collaborativemanagement? Public Adm. Rev. 66, 111–121 (2006).doi: 10.1111/j.1540-6210.2006.00671.x

20. J. Munck af Rosenschöld, N. Honkela, J. I. Hukkinen,Addressing the temporal fit of institutions: The regulation ofendocrine-disrupting chemicals in Europe. Ecol. Soc. 19, 30(2014). doi: 10.5751/ES-07033-190430

21. J. Raab, R. S. Mannak, B. Cambre, Combining structure,governance, and context: A configurational approach tonetwork effectiveness. J. Public Adm. Res. Theory 25,479–511 (2015). doi: 10.1093/jopart/mut039

22. A. D. Henry, M. Lubell, M. McCoy, Belief systems andsocial capital as drivers of policy network structure:The case of California regional planning. J. Public Adm.Res. Theory 21, 419–444 (2011). doi: 10.1093/jopart/muq042

23. E.-H. Klijn, C. Skelcher, Democracy and governance networks:Compatible or not? Public Adm. 85, 587–608 (2007).doi: 10.1111/j.1467-9299.2007.00662.x

24. F. R. Westley et al., A theory of transformative agency inlinked social-ecological systems. Ecol. Soc. 18, 27 (2013).doi: 10.5751/ES-05072-180327

25. E.-H. Klijn, B. Steijn, J. Edelenbos, The impact of networkmanagement on outcomes in governance networks.Public Adm. 88, 1063–1082 (2010). doi: 10.1111/j.1467-9299.2010.01826.x

26. M. Shellenberger, T. Nordhaus, The death ofenvironmentalism, Geopolit. Hist. Int. Relations 1, 121–163(2009).

27. A. Zachrisson, K. Beland Lindahl, Conflict resolution throughcollaboration: Preconditions and limitations in forest andnature conservation controversies. For. Policy Econ. 33,39–46 (2013). doi: 10.1016/j.forpol.2013.04.008

28. T. A. Muñoz-Erickson et al., Spanning boundaries in anArizona watershed partnership: Information networks as toolsfor entrenchment or ties for collaboration? Ecol. Soc. 15, 22(2010). doi: 10.5751/ES-03390-150322

29. A. P. Castro, E. Nielsen, Indigenous people andco-management: Implications for conflict management.Environ. Sci. Policy 4, 229–239 (2001). doi: 10.1016/S1462-9011(01)00022-3

30. P. A. Walker, P. T. Hurley, Collaboration derailed: The politicsof “community-based” resource management in NevadaCounty. Soc. Nat. Resour. 17, 735–751 (2004). doi: 10.1080/08941920490480723

31. R. F. Brummel, K. C. Nelson, P. J. Jakes, Burning throughorganizational boundaries? Examining inter-organizationalcommunication networks in policy-mandated collaborativebushfire planning groups. Glob. Environ. Change 22, 516–528(2012). doi: 10.1016/j.gloenvcha.2011.12.004

32. T. A. Scott, C. W. Thomas, Unpacking the collaborativetoolbox: Why and when do public managers choosecollaborative governance strategies? Policy Stud. J. 45,191–214 (2016). doi: 10.1111/psj.12162

33. C. Ansell, A. Gash, Collaborative governance in theory andpractice. J. Public Adm. Res. Theory 18, 543–571 (2007).doi: 10.1093/jopart/mum032

34. R. R. J. McAllister, R. McCrea, M. N. Lubell, Policy networks,stakeholder interactions and climate adaptation in the region ofSouth East Queensland, Australia. Reg. Environ. Change 14,527–539 (2014). doi: 10.1007/s10113-013-0489-4

35. D. Nohrstedt, Ö. Bodin, Evolutionary dynamics of crisispreparedness collaboration: Resources, turbulence andnetwork change in Swedish municipalities. Public Policy 5,134–155 (2014). doi: 10.1002/rhc3.12055

36. B. Nykvist, Does social learning lead to better naturalresource management? A case study of the modernfarming community of practice in Sweden. Soc. Nat.Resour. 27, 436–450 (2014). 10.1080/08941920.2013.861562

37. P. Matous, Y. Todo, Exploring dynamic mechanisms oflearning networks for resource conservation. Ecol. Soc. 20,36 (2015). doi: 10.5751/ES-07602-200236

38. A. D. Henry, B. Vollan, Networks and the challenge ofsustainable development. Annu. Rev. Environ. Resour. 39,583–610 (2014). doi: 10.1146/annurev-environ-101813-013246

39. Ö. Bodin, B. I. Crona, The role of social networks in naturalresource governance: What relational patterns make adifference? Glob. Environ. Change 19, 366–374 (2009).doi: 10.1016/j.gloenvcha.2009.05.002

40. C. S. Holling, G. K. Meffe, Command and control and thepathology of natural resource management. Conserv. Biol. 10,328–337 (1996). doi: 10.1046/j.1523-1739.1996.10020328.x

41. C. Folke, L. Pritchard Jr., F. Berkes, J. Colding, U. Svedin, Theproblem of fit between ecosystems and institutions: Tenyears later. Ecol. Soc. 12, 30 (2007). doi: 10.5751/ES-02064-120130

42. J. S. Sayles, J. A. Baggio, Social-ecological network analysisof scale mismatches in estuary watershed restoration.Proc. Natl. Acad. Sci. U.S.A. 114, E1776–E1785 (2017).doi: 10.1073/pnas.1604405114; pmid: 28223529

43. Ö. Bodin et al., Theorizing benefits and constraints incollaborative environmental governance: A transdisciplinarysocial-ecological network approach for empiricalinvestigations. Ecol. Soc. 21, 40 (2016). doi: 10.5751/ES-08368-210140

44. L. E. Dee et al., Operationalizing network theory forecosystem service assessments. Trends Ecol. Evol. 32,118–130 (2017). doi: 10.1016/j.tree.2016.10.011;pmid: 27856059

45. J. A. Ekstrom, O. R. Young, Evaluating functional fit betweena set of institutions and an ecosystem. Ecol. Soc. 14, 16(2009). doi: 10.5751/ES-02930-140216

46. K. Ingold, Network structures within policy processes:Coalitions, power, and brokerage in Swiss climate policy.Policy Stud. J. 39, 435–459 (2011). doi: 10.1111/j.1541-0072.2011.00416.x

47. C. T. Gallemore, R. D. Prasti H, M. Moeliono, Discursivebarriers and cross-scale forest governance in CentralKalimantan, Indonesia. Ecol. Soc. 19, 18 (2014). doi: 10.5751/ES-06418-190218

48. L. Jasny, J. Waggle, D. R. Fisher, An empirical examination ofecho chambers in US climate policy networks. Nat. Clim.Chang. 5, 782–786 (2015). doi: 10.1038/nclimate2666

49. M. L. Barnes, J. Lynham, K. Kalberg, P. Leung, Socialnetworks and environmental outcomes. Proc. Natl. Acad.Sci. U.S.A. 113, 6466–6471 (2016). doi: 10.1073/pnas.1523245113; pmid: 27217551

50. Ö. Bodin, A. Sandström, B. Crona, Collaborative networks foreffective ecosystem-based management: A set of workinghypotheses. Policy Stud. J. 45, 289–314 (2017). doi: 10.1111/psj.12146

51. T. W. Valente, Network interventions. Science 337, 49–53(2012). doi: 10.1126/science.1217330; pmid: 22767921

52. M. E. Isaac, Agricultural information exchange andorganizational ties: The effect of network topology onmanaging agrodiversity. Agric. Syst. 109, 9–15 (2012).doi: 10.1016/j.agsy.2012.01.011

53. F. Westley et al., Tipping toward sustainability: Emergingpathways of transformation. Ambio 40, 762–780 (2011).pmid: 22338714

54. B. Uzzi, J. Spiro, Collaboration and creativity: The small worldproblem. Am. J. Sociol. 111, 447–504 (2005). doi: 10.1086/432782

55. A.-M. Dowd et al., The role of networks in transformingAustralian agriculture. Nat. Clim. Chang. 4, 558–563 (2014).doi: 10.1038/nclimate2275

56. M. Granovetter, The strength of weak ties. Am. J. Sociol. 78,1360–1380 (1973). doi: 10.1086/225469

57. R. Berardo, The evolution of self-organizing communicationnetworks in high-risk social-ecological systems. Int. J.Commons 8, 236–258 (2014). doi: 10.18352/ijc.463

58. M. Lubell, L. Jasny, A. Hastings, Network governance forinvasive species management, Conserv. Lett. 10.1111/conl.12311 (2016).

59. R. Berardo, J. T. Scholz, Self-organizing policy networks: Risk,partner selection, and cooperation in estuaries. Am. J.Pol. Sci. 54, 632–649 (2010). doi: 10.1111/j.1540-5907.2010.00451.x

60. R. R. J. McAllister, B. M. Taylor, B. P. Harman, Partnershipnetworks for urban development: How structure is shaped byrisk. Policy Stud. J. 43, 379–398 (2015). doi: 10.1111/psj.12103




Dow

nloaded from

http://dx.doi.org/10.1111/psj.12028http://dx.doi.org/10.1016/j.cosust.2015.03.005http://dx.doi.org/10.5751/ES-01569-110114http://dx.doi.org/10.1016/j.cosust.2014.08.009http://dx.doi.org/10.5751/ES-05093-180103http://dx.doi.org/10.1126/science.1175325http://dx.doi.org/10.1126/science.1175325http://www.ncbi.nlm.nih.gov/pubmed/19745137http://dx.doi.org/10.2307/2269460http://dx.doi.org/10.1146/annurev.energy.30.050504.144511http://dx.doi.org/10.1146/annurev.energy.30.050504.144511http://dx.doi.org/10.5751/ES-03564-1504r01http://dx.doi.org/10.1007/s13280-014-0501-3http://dx.doi.org/10.1007/s13280-014-0501-3http://www.ncbi.nlm.nih.gov/pubmed/24659474http://dx.doi.org/10.1080/14735903.2012.751701http://dx.doi.org/10.1007/BF00136406http://dx.doi.org/10.1111/psj.12064http://dx.doi.org/10.1007/s11077-013-9181-2http://dx.doi.org/10.1111/j.1523-1739.2012.01850.xhttp://dx.doi.org/10.1111/j.1523-1739.2012.01850.xhttp://www.ncbi.nlm.nih.gov/pubmed/22624623http://dx.doi.org/10.1126/science.1091015http://dx.doi.org/10.1126/science.1091015http://www.ncbi.nlm.nih.gov/pubmed/14671286http://dx.doi.org/10.1111/j.1540-6210.2006.00671.xhttp://dx.doi.org/10.5751/ES-07033-190430http://dx.doi.org/10.1093/jopart/mut039http://dx.doi.org/10.1093/jopart/muq042http://dx.doi.org/10.1093/jopart/muq042http://dx.doi.org/10.1111/j.1467-9299.2007.00662.xhttp://dx.doi.org/10.5751/ES-05072-180327http://dx.doi.org/10.1111/j.1467-9299.2010.01826.xhttp://dx.doi.org/10.1111/j.1467-9299.2010.01826.xhttp://dx.doi.org/10.1016/j.forpol.2013.04.008http://dx.doi.org/10.5751/ES-03390-150322http://dx.doi.org/10.1016/S1462-9011(01)00022-3http://dx.doi.org/10.1016/S1462-9011(01)00022-3http://dx.doi.org/10.1080/08941920490480723http://dx.doi.org/10.1080/08941920490480723http://dx.doi.org/10.1016/j.gloenvcha.2011.12.004http://dx.doi.org/10.1111/psj.12162http://dx.doi.org/10.1093/jopart/mum032http://dx.doi.org/10.1007/s10113-013-0489-4http://dx.doi.org/10.1002/rhc3.12055http://10.1080/08941920.2013.861562http://10.1080/08941920.2013.861562http://dx.doi.org/10.5751/ES-07602-200236http://dx.doi.org/10.1146/annurev-environ-101813-013246http://dx.doi.org/10.1146/annurev-environ-101813-013246http://dx.doi.org/10.1016/j.gloenvcha.2009.05.002http://dx.doi.org/10.1046/j.1523-1739.1996.10020328.xhttp://dx.doi.org/10.5751/ES-02064-120130http://dx.doi.org/10.5751/ES-02064-120130http://dx.doi.org/10.1073/pnas.1604405114http://www.ncbi.nlm.nih.gov/pubmed/28223529http://dx.doi.org/10.5751/ES-08368-210140http://dx.doi.org/10.5751/ES-08368-210140http://dx.doi.org/10.1016/j.tree.2016.10.011http://www.ncbi.nlm.nih.gov/pubmed/27856059http://dx.doi.org/10.5751/ES-02930-140216http://dx.doi.org/10.1111/j.1541-0072.2011.00416.xhttp://dx.doi.org/10.1111/j.1541-0072.2011.00416.xhttp://dx.doi.org/10.5751/ES-06418-190218http://dx.doi.org/10.5751/ES-06418-190218http://dx.doi.org/10.1038/nclimate2666http://dx.doi.org/10.1073/pnas.1523245113http://dx.doi.org/10.1073/pnas.1523245113http://www.ncbi.nlm.nih.gov/pubmed/27217551http://dx.doi.org/10.1111/psj.12146http://dx.doi.org/10.1111/psj.12146http://dx.doi.org/10.1126/science.1217330http://www.ncbi.nlm.nih.gov/pubmed/22767921http://dx.doi.org/10.1016/j.agsy.2012.01.011http://www.ncbi.nlm.nih.gov/pubmed/22338714http://dx.doi.org/10.1086/432782http://dx.doi.org/10.1086/432782http://dx.doi.org/10.1038/nclimate2275http://dx.doi.org/10.1086/225469http://dx.doi.org/10.18352/ijc.463http://dx.doi.org/10.1111/j.1540-5907.2010.00451.xhttp://dx.doi.org/10.1111/j.1540-5907.2010.00451.xhttp://dx.doi.org/10.1111/psj.12103http://science.sciencemag.org/

61. R. Grundmann, Climate change as a wicked social problem.Nat. Geosci. 9, 562–563 (2016). doi: 10.1038/ngeo2780

62. C. L. Meek, Forms of collaboration and social fit in wildlifemanagement: A comparison of policy networks in Alaska.Glob. Environ. Change 23, 217–228 (2013). doi: 10.1016/j.gloenvcha.2012.10.003

63. A. Boin, P. ’t Hart, Organising for effective emergencymanagement: Lessons from research. Aust. J.Public Adm. 69, 357–371 (2010). doi: 10.1111/j.1467-8500.2010.00694.x

64. R. Biggs, M. Schlüter, M. L. Schoon, Principles for BuildingResilience: Sustaining Ecosystem Services in Social-EcologicalSystems (Cambridge Univ. Press, 2014).

65. I. Chadès et al., General rules for managing and surveyingnetworks of pests, diseases, and endangered species.Proc. Natl. Acad. Sci. U.S.A. 108, 8323–8328 (2011).doi: 10.1073/pnas.1016846108; pmid: 21536884

66. J. S. Lansing, J. N. Kremer, Emergent properties of Balinesewater temple networks: Coadaptation on a rugged fitnesslandscape. Am. Anthropol. 95, 97–114 (1993). doi: 10.1525/aa.1993.95.1.02a00050

67. P. Wang, G. Robins, P. Pattison, E. Lazega, Exponentialrandom graph models for multilevel networks. Soc. Networks35, 96–115 (2013). doi: 10.1016/j.socnet.2013.01.004

68. G. Hardin, The tragedy of the commons. Science 162,1243–1248 (1968). doi: 10.1126/science.162.3859.1243;pmid: 5699198

69. E. Ostrom, Governing the Commons: The Evolution of Institutionsfor Collective Action (Cambridge Univ. Press, 1990).

70. Ö. Bodin, B. Crona, M. Thyresson, A.-L. Golz, M. Tengö,Conservation success as a function of good alignment ofsocial and ecological structures and processes. Conserv. Biol.28, 1371–1379 (2014). doi: 10.1111/cobi.12306

71. A. M. Guerrero, Ö. Bodin, R. R. J. McAllister, K. A. Wilson,Achieving social-ecological fit through bottom-upcollaborative governance: An empirical investigation.Ecol. Soc. 20, 41 (2015). doi: 10.5751/ES-08035-200441

72. E. Treml, P. I. J. Fidelman, S. Kininmonth, J. Ekstrom,Ö. Bodin, Analyzing the (mis)fit between the institutional andecological networks of the Indo-West Pacific. Glob. Environ.Change 31, 263–271 (2015). doi: 10.1016/j.gloenvcha.2015.01.012

73. A. Bergsten, D. Galafassi, Ö. Bodin, The problem of spatial fitin social-ecological systems: Detecting mismatches betweenecological connectivity and land management in an urbanregion. Ecol. Soc. 19, 6 (2014). doi: 10.5751/ES-06931-190406

74. A. Bergsten, A. Zetterberg, To model the landscape as anetwork: A practitioner’s perspective. Landsc. Urban Plan.119, 35–43 (2013). doi: 10.1016/j.landurbplan.2013.06.009

75. M. Dallimer, N. Strange, Why socio-political borders andboundaries matter in conservation. Trends Ecol. Evol. 30,132–139 (2015). doi: 10.1016/j.tree.2014.12.004;pmid: 25637933

76. S. Kininmonth, A. Bergsten, Ö. Bodin, Closing thecollaborative gap: Aligning social and ecological connectivityfor better management of interconnected wetlands. Ambio44, S138–S148 (2015). doi: 10.1007/s13280-014-0605-9;pmid: 25576288

77. E. Ostrom, Polycentric systems for coping with collectiveaction and global environmental change. Glob. Environ.Change 20, 550–557 (2010). doi: 10.1016/j.gloenvcha.2010.07.004

78. C. Wyborn, R. P. Bixler, Collaboration and nestedenvironmental governance: Scale dependency, scale framing,and cross-scale interactions in collaborative conservation.J. Environ. Manage. 123, 58–67 (2013). doi: 10.1016/j.jenvman.2013.03.014; pmid: 23583866

79. K. Maciejewski, A. De Vos, G. S. Cumming, C. Moore, D. Biggs,Cross-scale feedbacks and scale mismatches as influenceson cultural services and the resilience of protected areas.Ecol. Appl. 25, 11–23 (2015). doi: 10.1890/13-2240.1;pmid: 26255354

80. Ö. Bodin, D. Nohrstedt, Formation and performance ofcollaborative disaster management networks: Evidence froma Swedish wildfire response. Glob. Environ. Change 41,183–194 (2016). doi: 10.1016/j.gloenvcha.2016.10.004

81. M. Lubell, G. Robins, P. Wang, Network structure andinstitutional complexity in an ecology of water managementgames. Ecol. Soc. 19, 23 (2014). doi: 10.5751/ES-06880-190423

82. R. R. J. McAllister et al., Balancing collaboration withcoordination: Contesting eradication in the Australian plantpest and disease biosecurity system. Int. J. Commons 11, 330(2017). doi: 10.18352/ijc.701

83. K. C. Nelson, R. F. Brummel, N. Jordan, S. Manson, Socialnetworks in complex human and natural systems: The case ofrotational grazing, weak ties, and eastern US dairylandscapes. Agric. Human Values 31, 245–259 (2014).doi: 10.1007/s10460-013-9462-6

84. K. Ingold, M. Fischer, Drivers of collaboration to mitigateclimate change: An illustration of Swiss climate policy over 15years. Glob. Environ. Change 24, 88–98 (2014). doi: 10.1016/j.gloenvcha.2013.11.021

85. S. M. Alexander, D. Armitage, P. J. Carrington, Ö. Bodin,Examining horizontal and vertical social ties to achievesocial-ecological fit in an emerging marine reserve network.Aquat. Conserv. 10.1002/aqc.2775 (2017).

86. B. Crona, S. Gelcich, Ö. Bodin, The importance of interplaybetween leadership and social capital in shaping outcomes ofrights-based fisheries governance. World Dev. 91, 70–83(2017). doi: 10.1016/j.worlddev.2016.10.006

87. S. Alexander, D. Armitage, T. Charles, Social networks andtransitions to co-management in Jamaican marine reservesand small-scale fisheries. Glob. Environ. Change 35, 213–225(2015). doi: 10.1016/j.gloenvcha.2015.09.001

88. R. S. Burt, Structural holes and good ideas. Am. J. Sociol. 110,349–399 (2004). doi: 10.1086/421787

89. P. Olsson, V. Galaz, W. J. Boonstra, Sustainabilitytransformations: A resilience perspective. Ecol. Soc. 19,1 (2014). doi: 10.5751/ES-06799-190401

90. M. Barnes, K. Kalberg, M. Pan, P. Leung, When is brokeragenegatively associated with economic benefits? Ethnicdiversity, competition, and common-pool resources.Soc. Networks 45, 55–65 (2016). doi: 10.1016/j.socnet.2015.11.004

91. A. Sandström, C. Rova, Adaptive co-management networks:A comparative analysis of two fishery conservation areas inSweden. Ecol. Soc. 15, 14 (2010). doi: 10.5751/ES-03531-150314

92. J. Shore, E. Bernstein, D. Lazer, Facts and figuring: Anexperimental investigation of network structure andperformance in information and solution spaces. Organ. Sci.26, 1432–1446 (2015). doi: 10.1287/orsc.2015.0980

93. K. G. Provan, P. Kenis, Modes of network governance:Structure, management, and effectiveness. J. Public Adm.Res. Theory 18, 229–252 (2007). doi: 10.1093/jopart/mum015

94. K. Vance-Borland, J. Holley, Conservation stakeholdernetwork mapping, analysis, and weaving. Conserv. Lett. 4,278–288 (2011). doi: 10.1111/j.1755-263X.2011.00176.x

95. D. G. Rand, S. Arbesman, N. A. Christakis, Dynamic socialnetworks promote cooperation in experiments with humans.Proc. Natl. Acad. Sci. U.S.A. 108, 19193–19198 (2011).doi: 10.1073/pnas.1108243108; pmid: 22084103

96. R. Berardo, M. Lubell, Understanding what shapes apolycentric governance system. Public Adm. Rev. 76,738–751 (2016). doi: 10.1111/puar.12532

97. M. D. McGinnis, Networks of adjacent action situations inpolycentric governance. Policy Stud. J. 39, 51–78 (2011).doi: 10.1111/j.1541-0072.2010.00396.x

98. M. Lubell, J. M. Mewhirter, R. Berardo, J. T. Scholz,Transaction costs and the perceived effectiveness of complexinstitutional systems. Public Adm. Rev. 10.1111/puar.12622(2016).

99. T. A. Scott, C. W. Thomas, Winners and losers in the ecologyof games: Network position, connectivity, and the benefits ofcollaborative governance regimes. J. Public Adm. Res. Theory31, 59 (2017). doi: 10.1093/jopart/mux009

100. R. Plummer et al., Is adaptive co-management delivering?Examining relationships between collaboration, learning andoutcomes in UNESCO biosphere reserves. Ecol. Econ. 140,79–88 (2017). doi: 10.1016/j.ecolecon.2017.04.028

ACKNOWLEDGMENTS

This work was financially supported by the Swedish Foundation forStrategic Environmental Research (Mistra) through a core grant tothe Stockholm Resilience Centre at Stockholm University. TheSwedish Research Council and Formas contributed with additionalsupport through project grants 2016-04263 and 2016-01137.

10.1126/science.aan1114




Dow

nloaded from

http://dx.doi.org/10.1038/ngeo2780http://dx.doi.org/10.1016/j.gloenvcha.2012.10.003http://dx.doi.org/10.1016/j.gloenvcha.2012.10.003http://dx.doi.org/10.1111/j.1467-8500.2010.00694.xhttp://dx.doi.org/10.1111/j.1467-8500.2010.00694.xhttp://dx.doi.org/10.1073/pnas.1016846108http://www.ncbi.nlm.nih.gov/pubmed/21536884http://dx.doi.org/10.1525/aa.1993.95.1.02a00050http://dx.doi.org/10.1525/aa.1993.95.1.02a00050http://dx.doi.org/10.1016/j.socnet.2013.01.004http://dx.doi.org/10.1126/science.162.3859.1243http://www.ncbi.nlm.nih.gov/pubmed/5699198http://dx.doi.org/10.1111/cobi.12306http://dx.doi.org/10.5751/ES-08035-200441http://dx.doi.org/10.1016/j.gloenvcha.2015.01.012http://dx.doi.org/10.1016/j.gloenvcha.2015.01.012http://dx.doi.org/10.5751/ES-06931-190406http://dx.doi.org/10.5751/ES-06931-190406http://dx.doi.org/10.1016/j.landurbplan.2013.06.009http://dx.doi.org/10.1016/j.tree.2014.12.004http://www.ncbi.nlm.nih.gov/pubmed/25637933http://dx.doi.org/10.1007/s13280-014-0605-9http://www.ncbi.nlm.nih.gov/pubmed/25576288http://dx.doi.org/10.1016/j.gloenvcha.2010.07.004http://dx.doi.org/10.1016/j.gloenvcha.2010.07.004http://dx.doi.org/10.1016/j.jenvman.2013.03.014http://dx.doi.org/10.1016/j.jenvman.2013.03.014http://www.ncbi.nlm.nih.gov/pubmed/23583866http://dx.doi.org/10.1890/13-2240.1http://www.ncbi.nlm.nih.gov/pubmed/26255354http://dx.doi.org/10.1016/j.gloenvcha.2016.10.004http://dx.doi.org/10.5751/ES-06880-190423http://dx.doi.org/10.5751/ES-06880-190423http://dx.doi.org/10.18352/ijc.701http://dx.doi.org/10.1007/s10460-013-9462-6http://dx.doi.org/10.1016/j.gloenvcha.2013.11.021http://dx.doi.org/10.1016/j.gloenvcha.2013.11.021http://dx.doi.org/10.1016/j.worlddev.2016.10.006http://dx.doi.org/10.1016/j.gloenvcha.2015.09.001http://dx.doi.org/10.1086/421787http://dx.doi.org/10.5751/ES-06799-190401http://dx.doi.org/10.1016/j.socnet.2015.11.004http://dx.doi.org/10.1016/j.socnet.2015.11.004http://dx.doi.org/10.5751/ES-03531-150314http://dx.doi.org/10.5751/ES-03531-150314http://dx.doi.org/10.1287/orsc.2015.0980http://dx.doi.org/10.1093/jopart/mum015http://dx.doi.org/10.1093/jopart/mum015http://dx.doi.org/10.1111/j.1755-263X.2011.00176.xhttp://dx.doi.org/10.1073/pnas.1108243108http://www.ncbi.nlm.nih.gov/pubmed/22084103http://dx.doi.org/10.1111/puar.12532http://dx.doi.org/10.1111/j.1541-0072.2010.00396.xhttp://dx.doi.org/10.1093/jopart/mux009 http://dx.doi.org/10.1016/j.ecolecon.2017.04.028http://science.sciencemag.org/

systemsCollaborative environmental governance: Achieving collective action in social-ecological

Örjan Bodin

DOI: 10.1126/science.aan1114 (6352), eaan1114.357Science

, this issue p. eaan1114Scienceneeded.governance and points out that there remain substantial knowledge gaps and key areas where more research is

conclusions about the benefits and constraints of collaborative approaches to environmental management andand what kind of environmental problems are most fruitfully addressed in this way. The piece provides general environmental problems. Bodin reviews studies and cases that elucidate when, if, and how collaboration can be effectiveof stakeholders often lack the willingness to deliberate and contribute to jointly negotiated solutions to common

By its nature, environmental governance requires collaboration. However, studies have shown that various typesCollaborative governance

ARTICLE TOOLS http://science.sciencemag.org/content/357/6352/eaan1114

REFERENCES

http://science.sciencemag.org/content/357/6352/eaan1114#BIBLThis article cites 93 articles, 8 of which you can access for free

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAAS.ScienceScience, 1200 New York Avenue NW, Washington, DC 20005. The title (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement ofScience

Science. No claim to original U.S. Government WorksCopyright © 2017 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of

on July 2, 2021


Dow

nloaded from
http://science.sciencemag.org/content/357/6352/eaan1114http://science.sciencemag.org/content/357/6352/eaan1114#BIBLhttp://www.sciencemag.org/help/reprints-and-permissionshttp://www.sciencemag.org/about/terms-servicehttp://science.sciencemag.org/

environment collaborative environmental governance ...Örjan bodin* managing ecosystems is...

Documents